An ecological approach to preventing human infection: vaccinating wild mouse reservoirs intervenes in the Lyme disease cycle

Abstract

Many pathogens, such as the agents of West Nile encephalitis and plague, are maintained in nature by animal reservoirs and transmitted to humans by arthropod vectors. Efforts to reduce disease incidence usually rely on vector control or immunization of humans. Lyme disease, for which no human vaccine is currently available, is a commonly reported vector-borne disease in North America and Europe. In a recently developed, ecological approach to disease prevention, we intervened in the natural cycle of the Lyme disease agent (Borrelia burgdorferi) by immunizing wild white-footed mice (Peromyscus leucopus), a reservoir host species, with either a recombinant antigen of the pathogen, outer surface protein A, or a negative control antigen in a repeated field experiment with paired experimental and control grids stratified by site. Outer surface protein A vaccination significantly reduced the prevalence of B. burgdorferi in nymphal blacklegged ticks (Ixodes scapularis) collected at the sites the following year in both experiments. The magnitude of the vaccine's effect at a given site correlated with the tick infection prevalence found on the control grid, which in turn correlated with mouse density. These data, as well as differences in the population structures of B. burgdorferi in sympatric ticks and mice, indicated that nonmouse hosts contributed more to infecting ticks than previously expected. Thus, where nonmouse hosts play a large role in infection dynamics, vaccination should be directed at additional species.

Fig. 1.

Anti-OspA antibody responses in recombinant OspA- and GST-immunized Peromyscus leucopus populations at a Connecticut field site in 1998 (A) and 2000 (B). The histograms with 95% confidence interval error bars indicate the antibody index by OspA enzyme immunoassay (left y axis) for each group of mice by treatment, site, and trapping period. The dots indicate the percentages of the serum samples in each group that exceeded the mean of negative control sera in the EIA by 3 standard deviations (SD; right y axis). The numbers of P. leucopus sampled each trapping period are given at the top of each graph.

Fig. 2.

Distribution of group I (A), II (B), and III (C) B. burgdorferi genotype groups among questing nymphs collected on GST control (open bars) and OspA vaccine (shaded bars) grids in 2002. Numbers of nymphs assayed per grid are listed within bars. OR and 95% CI given for the odds of a nymph being infected with a given genotype on a vaccinated versus a control grid were determined by an exact Cochran–Mantel–Haenzsel procedure for stratified tables. Some nymphs were infected with two B. burgdorferi genotype groups; thus, the sum of the prevalence of genotype groups was greater than the overall nymphal infection prevalence for some groups.

Fig. 3.

The relationship between mouse density, nymphal infection prevalence, and magnitude of vaccine effect. (A) The relationship between mouse density (mice/ha, hectare) and nymphal infection (inf.) prevalence. Dashed lines represent 95% CI of the slope for the solid regression lines. The coefficient of determination (r²) is given for each treatment. Open and filled symbols represent GST-immunized (control) and OspA-immunized grids, respectively. Grids from the same site are represented by the same symbol: site 1, circle; site 2, square; site 3, inverted triangle; site 4, hexagon; site 5, triangle; site 6, diamond. (B) Vaccine effect (1 – OR) as a function of the control grid nymphal infection prevalence (Table 1). The OR represents the odds of a nymph being infected on a vaccinated grid versus on the control grid at the same site. Symbols for sites are as indicated in A.